Characterization and prediction of meandering channel migration in the GIS environment: a case study of the Sabine River in the USA.

This study focused on the prediction of a 22 km meandering channel migration of the Sabine River between the states of Texas and Louisiana. The meander characteristics of 12 bends, identified from seven orthophotos taken between 1974 and 2004, were acquired in a GIS environment. Based on that earlier years' data acquisition, channel prediction was performed for the two years 1996 and 2004 using least squares estimation and linear extrapolations, yielding a satisfactory agreement with the observations (the median predicted and observed migration rates were 3.1 and 3.6 [m/year], respectively). The best-predicted migration rate was found to be associated with the longest orthophoto-recorded interval. The study confirmed that channel migration is strongly correlated with bend curvature and that the maximum migration rate of the bend corresponded to a radius of curvature [bend radius (R(C))/channel width (W(C))] of 2.5. In tight bends of a smaller radius of curvature than 1.6, secondary flow scouring near the bend apex increases bend curvature. The stability index of the dimensionless bend radius was determined to be 2.45. Overall, this study proves the effectiveness of least squares estimation with historical orthophotography for characterization of meandering channel migration.